core: stage 9 — rules/ (filename sanitization, collision policy, rule matching)

Pure functions only, per AGENT-CORE.md's build order: no I/O, no JSON, no SQL,
no notion of the wire Rule type — DAEMON decodes its own stored/wire
representation into these plain structs and calls in.

- rules/filename.hpp: sanitize_filename() turns a raw candidate (from
  net::parse_content_disposition or net::url_filename — neither is
  filesystem-safe by design; both headers say so and point here) into one
  safe to create on ext4/APFS/NTFS: strips separators and control bytes,
  folds NTFS-illegal characters, neutralizes reserved Windows device names,
  clamps length on a UTF-8 boundary. Total on hostile input; never empty.
  Not the path-traversal security boundary — that's daemon/fs/safepath,
  downstream of this and the one that actually matters adversarially.
- rules/collision.hpp: resolve_collision() finds the next free name
  Explorer/Finder-style ("name (1).ext", ...) given an existence predicate,
  or returns the desired name unchanged under an overwrite policy. Never
  fabricates a guaranteed-unique name past its attempt bound — hands back
  the last candidate tried rather than hiding a persistent collision.
- rules/match.hpp: match_rules() is the evaluation half of
  contracts/schema/types/Rule.schema.json — priority order, first rule
  whose present match clauses (extensions/mimeTypes/host & url glob/size
  bounds) all hold, wins; a size clause never matches speculatively before
  the probe fills in size_bytes. glob_match() is the iterative (not
  recursive — bounded work on an all-'*' pattern) matcher both host_pattern
  and url_pattern use.

Every header compiles standalone; tests (39 cases) pass under ASan+UBSan and
TSan. core/include/vdm/README.md documents the new public surface.

Co-Authored-By: Claude Sonnet 5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01Q3QrF7rCt21bkAjt9BCDFQ

core/src/rules/match.cpp

@@ -0,0 +1,98 @@
+#include "vdm/rules/match.hpp"
+
+#include <algorithm>
+#include <cctype>
+#include <cstddef>
+#include <numeric>
+
+namespace vdm::rules {
+namespace {
+
+char lower_ascii(char c) { return static_cast<char>(std::tolower(static_cast<unsigned char>(c))); }
+
+bool ieq(std::string_view a, std::string_view b) {
+    return a.size() == b.size() &&
+           std::equal(a.begin(), a.end(), b.begin(),
+                     [](char x, char y) { return lower_ascii(x) == lower_ascii(y); });
+}
+
+// A rule's `extensions` entries are documented with no leading '.', but be lenient about
+// one showing up anyway (a hand-edited rule table, an older client) rather than let a
+// clause that never matches silently swallow a whole category.
+std::string_view strip_leading_dot(std::string_view s) {
+    return (!s.empty() && s.front() == '.') ? s.substr(1) : s;
+}
+
+bool match_clause(const RuleMatch &m, const MatchInput &in) {
+    if (m.extensions) {
+        bool any = std::any_of(m.extensions->begin(), m.extensions->end(), [&](const auto &e) {
+            return ieq(strip_leading_dot(e), in.extension);
+        });
+        if (!any)
+            return false;
+    }
+    if (m.mime_types) {
+        bool any = std::any_of(m.mime_types->begin(), m.mime_types->end(),
+                               [&](const auto &t) { return ieq(t, in.mime_type); });
+        if (!any)
+            return false;
+    }
+    if (m.host_pattern && !glob_match(*m.host_pattern, in.host))
+        return false;
+    if (m.url_pattern && !glob_match(*m.url_pattern, in.url))
+        return false;
+    if (m.min_size_bytes) {
+        if (!in.size_bytes || *in.size_bytes < *m.min_size_bytes)
+            return false;
+    }
+    if (m.max_size_bytes) {
+        if (!in.size_bytes || *in.size_bytes > *m.max_size_bytes)
+            return false;
+    }
+    return true;
+}
+
+}  // namespace
+
+bool glob_match(std::string_view pattern, std::string_view text) noexcept {
+    // Classic iterative wildcard match (single backtrack point at the most recent '*'), not
+    // the naive recursive version — bounded work on any input, including a pattern that is
+    // nothing but repeated '*'s against a long `text`.
+    std::size_t p = 0, t = 0;
+    std::size_t star = std::string_view::npos, mark = 0;
+    while (t < text.size()) {
+        if (p < pattern.size() && (pattern[p] == '?' || lower_ascii(pattern[p]) == lower_ascii(text[t]))) {
+            ++p;
+            ++t;
+        } else if (p < pattern.size() && pattern[p] == '*') {
+            star = p++;
+            mark = t;
+        } else if (star != std::string_view::npos) {
+            p = star + 1;
+            t = ++mark;
+        } else {
+            return false;
+        }
+    }
+    while (p < pattern.size() && pattern[p] == '*') ++p;
+    return p == pattern.size();
+}
+
+std::optional<RuleAction> match_rules(const std::vector<Rule> &rules, const MatchInput &input) {
+    std::vector<std::size_t> order(rules.size());
+    std::iota(order.begin(), order.end(), 0);
+    // Stable by construction: std::stable_sort keeps table order among equal priorities.
+    std::stable_sort(order.begin(), order.end(), [&](std::size_t a, std::size_t b) {
+        return rules[a].priority < rules[b].priority;
+    });
+    for (auto i : order) {
+        const Rule &r = rules[i];
+        if (!r.enabled)
+            continue;
+        if (match_clause(r.match, input))
+            return r.action;
+    }
+    return std::nullopt;
+}
+
+}  // namespace vdm::rules